Contrastive Search in large language models addresses issues such as degenerative expressions and a lack of semantic coherence.
An article in The New York Times tells us: Microsoft Says New A.I. Shows Signs of Human Reasoning. MSFT researchers claim that GPT-4 solves puzzles and generates text with a deep and flexible understanding of human concepts and skills. Others
Generative AI and the Law: “Disgorgement” is a legal concept in which a court orders the return of any benefits gained from wrongful conduct. It is a form of equitable relief used to prevent unjust enrichment, restore the status quo
For those interested in plunging into the deep grass of how large language models and the like work, Damien Benveniste offers us a concise primer in Deep Neural Networks: All the Building Blocks. He gives us an overview of loss
Alex Kantowitz tells us that “Wacky, Unhinged Bing Chatbot Is Still Good For Microsoft’s Business.” The Bing chatbot has recently gone viral on social media with screenshots of it displaying unhinged behaviour and condescending responses. It gives false information, insults
Does ChatGPT Have a Terminal? Michael King asks ChatGPT to simulate a Linux terminal. Michael King tells us: I knew it! ChatGPT has Access to Internet — Linux Terminal Simulator is the Proof? Here’s the prompt he used: This text
Damien Benveniste provides us with two articles about the history and taxonomy of large language models. In a post on LinkedIn , he leads us from 2017’s Attention is all you need , to 2018’s ELMo and BERT. He references Facebooks XLM
The results showed that the open-source 6B parameter model exceeded the average few-shot performance of the GPT3-175B model on 15 out of 20 benchmarks. Jean-marc Mommessin informs us “Researchers developed a simple prompting strategy that enables open-source LLMs with 30x
ChatGPT and Education: Open AI has published a lengthy and detailed document on the ethical issues involving the use of generative AI in education. They discuss the many ways that this technology can be used as a tool for educators,
Astral Code Ten writes a lengthy and fascinating post exploring the true nature of AI entitled “Janus’ Simulators“. He notes that early (pre-2022) AI philosophers posited three types of AI: a goal-directed agent, a question answerer, and a direction follower.